New: 97,000+ skills indexed from GitHub repos with 100+ stars · every skill safety-checkedBrowse top skillsTop skills
Esc to closeBrowse all skills
Submit a skillBrowse skills
Submit a skillBrowse skills

Agent skill

Cxas Agent Foundry

by GoogleCloudPlatform in GoogleCloudPlatform/cxas-scrapi

End-to-end GECX/CXAS/CES conversational agent lifecycle -- build agents from requirements (PRD-to-agent), create and run evals (goldens, simulations, tool tests, callback tests), debug failures, and…

Apache-2.0Auto-check passedDevelopment

Install Cxas Agent Foundry

skills CLI
$ npx skills add GoogleCloudPlatform/cxas-scrapi --skill cxas-agent-foundry -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install GoogleCloudPlatform/cxas-scrapi cxas-agent-foundry --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/GoogleCloudPlatform/cxas-scrapi.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.agents/skills/cxas-agent-foundry .claude/skills/cxas-agent-foundry && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
cxas-agent-foundry
GitHub stars
106
Token cost
~2.4k tokens
SKILL.md length
809 words
Files
112 (incl. scripts, references, assets)
Skills in repo
14
Repo updated
First seen
Licence
Apache-2.0

At a glance

End-to-end GECX/CXAS/CES conversational agent lifecycle -- build agents from requirements (PRD-to-agent), create and run evals (goldens, simulations, tool tests, callback tests), debug failures, and…

  • Works in 3 steps: Virtualenv exists? -- Check if .venv/… → Config exists? -- Check if… → Has built before? -- Check if any…
  • The user mentions GECX
  • SKILL.md covers Step tracking — MANDATORY…, Quick Reference, Sub-agents and Environment Readiness Check…, plus 2 more sections
  • Runs Python scripts from its folder; calls python

What it does

Cxas Agent Foundry is an agent skill from GoogleCloudPlatform/cxas-scrapi. End-to-end GECX/CXAS/CES conversational agent lifecycle -- build agents from requirements (PRD-to-agent), create and run evals (goldens, simulations, tool tests, callback tests), debug failures, and iterate to production quality. Use this skill whenever the user mentions GECX, CXAS, CES, SCRAPI, conversational agents, voice agents, audio agents, agent evals, pushing/pulling/linting agents, or agent instructions/callbacks/tools on the Google Customer Engagement Suite platform.

Its SKILL.md is about 2.4k tokens, which your agent loads only when the skill is triggered. The skill folder holds 122 other files, including scripts, reference files and assets (for example `agents/coverage-analyst.md`, `agents/eval-writer.md` and `agents/lint-fixer.md`).

It sits in Development, covering LLM evaluation, Linting and formatting and PRD writing. The repository describes itself as: A powerful Python API, CLI, and set of Agent Skills for CX Agent Studio to automate, evaluate, and scale your agents with ease. The licence is Apache-2.0.

When your agent uses it

  • The user mentions GECX
  • Conversational agents
  • Pushing/pulling/linting agents
  • Agent instructions/callbacks/tools on the Google Customer Engagement Suite platform

Example prompts

  • “/cxas-agent-foundry”

Requirements

  • Python 3

Workflow steps

3 steps, taken from the first numbered list in SKILL.md.

  1. Virtualenv exists? -- Check if .venv/ directory exists
  2. Config exists? -- Check if .active-project file exists and the referenced /gecx-config.json exists
  3. Has built before? -- Check if any /cxas_app/ directory has content

What it can do on your machine

Read from SKILL.md and the folder at commit ffba639. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Ships 1 file in scripts/ (Python, from the files we listed), which the agent can run.

    Shell commands in SKILL.md call:

    • python

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Cxas Agent Foundry loads about 2.4k tokens when it runs, and up to ~52k if it reads all its reference files. Until then it costs about 125 tokens; SKILL.md has 809 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~125
When it runs · the whole SKILL.md, loaded when a task matches
~2.4k
With references · SKILL.md plus every file in references/, read only if the agent opens them
~52k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); the scripts in this folder are not scanned.

SKILL.md

The full file from GoogleCloudPlatform/cxas-scrapi at commit ffba639, republished under its Apache-2.0 licence (© GoogleCloudPlatform). 809 words, ~2,364 tokens.

Download SKILL.mdSave it as .claude/skills/cxas-agent-foundry/SKILL.md (or your agent's skills folder). This skill also uses 111 other files; get the full folder from GitHub.
name
cxas-agent-foundry
description
End-to-end GECX/CXAS/CES conversational agent lifecycle -- build agents from requirements (PRD-to-agent), create and run evals (goldens, simulations, tool tests, callback tests), debug failures, and iterate to production quality. Use this skill whenever the user mentions GECX, CXAS, CES, SCRAPI, conversational agents, voice agents, audio agents, agent evals, pushing/pulling/linting agents, or agent instructions/callbacks/tools on the Google Customer Engagement Suite platform.

Agent Foundry

End-to-end lifecycle for GECX conversational agents: build, test, debug, iterate.

Step tracking — MANDATORY (Phase 0, blocking)

Before doing ANY work — including running setup, asking questions, or scaffolding files — initialize <project>/todo.md from the relevant sub-skill's checklist (verbatim). The checklist is a contract, not a suggestion. If todo.md doesn't exist for the current task, refuse to proceed and create it first.

Long debug/build runs skip verification steps under pressure (e.g., pushing without linting, scaffolding without a TDD, claiming "deployed" without actually pushing). The checklist exists because of this. The instinct to skip a step is the moment the checklist earns its keep — that's when you must consult it, not the moment to bypass it.

Quick Reference

bash
# Lint: dispatch agents/lint-fixer.md sub-agent — DO NOT run `cxas lint` on the main thread.
# Lint output is verbose; keep it inside the sub-agent context.

# Push local files to platform (only after lint-fixer returns status: clean)
cxas push --app-dir <project>/cxas_app/<AppName> \
  --to projects/<project_id>/locations/<location>/apps/<app_id> \
  --project-id <project_id> --location <location>

# Pull platform state to local (use --version-id to export an immutable version snapshot instead of live draft)
cxas pull projects/<project_id>/locations/<location>/apps/<app_id> \
  --project-id <project_id> --location <location> --target-dir <project>/cxas_app/ \
  [--version-id <version_id_or_name>]

# Run evals + triage + report (single command)
python .agents/skills/cxas-agent-foundry/scripts/run-and-report.py --message "what changed" --runs 5

# Generate dynamic interactive HTML dashboard with Gemini LLM failure clustering and parameter filters
python .agents/skills/cxas-agent-foundry/scripts/generate_interactive_report.py --input <path_to_sim_results.json> --output <path_to_report.html>

# Inspect app architecture
python .agents/skills/cxas-agent-foundry/scripts/inspect-app.py

# Triage failures
python .agents/skills/cxas-agent-foundry/scripts/triage-results.py --last 3

# Run all 6 build-verification gates against the deployed app
python .agents/skills/cxas-agent-foundry/scripts/gate-check.py

# Tune scoring thresholds (similarity, hallucination, extra-tools)
python .agents/skills/cxas-agent-foundry/scripts/app-thresholds.py show

# Sync callback Python code into evals/callback_tests/agents/ + create test.py symlinks.
# Required for tests to be discoverable by test_all_callbacks_in_app_dir.
python .agents/skills/cxas-agent-foundry/scripts/sync-callbacks.py                  # post-push: pull from platform
python .agents/skills/cxas-agent-foundry/scripts/sync-callbacks.py --from-local <app_dir>  # pre-push: copy from local app dir

# Snapshot app version (create immutable platform backup)
cxas versions create --app-name projects/<project_id>/locations/<location>/apps/<app_id> \
  --display-name "v1.0.0-snapshot" --description "Pre-refactor baseline"

# List and compare app versions
cxas versions list --app-name projects/<project_id>/locations/<location>/apps/<app_id>
cxas versions compare --app-name projects/<project_id>/locations/<location>/apps/<app_id> \
  --source <version_id_1> --target <version_id_2> --web

# Cold-start setup (first-time only — venv + project bootstrap)
.agents/skills/cxas-agent-foundry/scripts/setup.sh
python .agents/skills/cxas-agent-foundry/scripts/setup-project.py

Disambiguation: gate-check.py and inspect-app.py overlap on "show me the architecture" but gate-check.py is the answer whenever the user is about to push, finished building, or wants a verification pass. inspect-app.py is for a quick "what's in here" look without the verification gates. When in doubt, use gate-check.py.

Sub-agents

For heavy diagnosis/analysis work that would otherwise burn main-thread context, dispatch one of these sub-agents via the Agent tool. Pass the contents of the relevant .md file as the prompt, then add the inputs the file lists.

Sub-agentReasoning intensityWhen to use
agents/triage-failure.mdHIGHDiagnose ONE failing eval. Fan out for the top 5 failures by category priority in parallel. Iterate on more after the first batch returns.
agents/tdd-writer.mdHIGHReverse-engineer a TDD from an existing agent OR draft from PRD. Returns the TDD + open-questions handoff; main thread runs the show/ask/iterate loop with the user (sub-agents can't ask).
agents/scaffolder.mdMEDIUMBulk-generate all agent code (agent JSONs, instruction.txt, tool python_code, callbacks, app.json) from an APPROVED TDD. One dispatch replaces 30-60 main-thread file writes.
agents/coverage-analyst.mdMEDIUMGenerate a full eval coverage report against an agent's architecture.
agents/eval-writer.mdMEDIUMGenerate evals for one entire eval TYPE (all goldens, all sims, etc.) — reads TDD's Coverage Map itself. Max 4 dispatches per build.
agents/lint-fixer.mdLOW (mechanical)Run cxas lint and mechanically fix all errors + deterministic warnings until clean. Never run lint on main thread.

For running evals: there is no sub-agent. Use scripts/run-and-report.py --json-summary <path> > /dev/null 2>&1 and read the summary file — see references/debug.md → "Quick Start". The work was deterministic, so it lives in the script.

Reasoning intensity is a hint to the runtime: HIGH sub-agents benefit from more thinking budget / a stronger model, LOW sub-agents are recipe-driven and don't. Each sub-agent file repeats this hint at the top with a one-line justification.

Environment Readiness Check (run BEFORE routing)

Before routing to any sub-skill, check these signals in order:

  1. Virtualenv exists? -- Check if .venv/ directory exists
  2. Config exists? -- Check if .active-project file exists and the referenced <project>/gecx-config.json exists
  3. Has built before? -- Check if any <project>/cxas_app/ directory has content
SignalAction
No .venv/ or no configFirst-time setup needed. Load references/setup.md before doing anything else.
gecx-config.json exists but no cxas_app/ contentReturning user, new project. Route normally.
All existReturning user. Route normally.
Show full SKILL.md (314 more words)Show less

Detect Intent and Route

Read what the user wants and load the appropriate sub-skill:

User says...PhaseLoad
"Build me an agent from this PRD"Buildreferences/build.md
"Create a new cxas app", "Make a new agent", "Set up an agent", "I wanna build an agent"Buildreferences/build.md
"Create evals for my agent"Buildreferences/build.md
"Generate tool tests", "create callback tests"Buildreferences/build.md
"Update evals -- requirements changed"Buildreferences/build.md
"Update the TDD"Buildreferences/build.md
"Run evals", "push evals", "check results"Runreferences/run.md
"Run tool tests", "test the callbacks"Runreferences/run.md
"Generate a report"Runreferences/run.md
"Generate interactive report", "cluster failures dashboard"Runreferences/generating-reports.md → "Interactive Diagnostic Dashboard"
"Why is this eval failing", "get to 90%"Debugreferences/debug.md
"Fix the failing evals", "debug the agent"Debugreferences/debug.md
"Tool test is failing", "callback test broke"Debugreferences/debug.md
"Snapshot app version", "Create version", "Compare versions", "List versions"Managereferences/api-reference.md → "Version Management"
"Edit the agent's instructions", "tweak the auth tool", "fix the greeting", "update this callback"Build (Edit cycle)references/build.md → "Editing an Existing Agent"

Any phrasing that implies creating, building, or setting up an agent/app routes to references/build.md — even if it sounds like "just create the app shell." "Create a new cxas app" is NOT a shortcut to scaffolding; it triggers the full build flow (todo.md → interview/PRD → TDD + approval → scaffold → lint → evals → push). Skipping the interview / TDD because the user said "create" instead of "build" is a routing failure.

Editing an existing agent (instruction tweak, tool change, callback fix) routes to build.md's "Editing an Existing Agent" section — the standard pull → edit → lint → push → run-evals cycle. Don't skip lint or the eval run after — silent regressions are how 90% rates drop to 70%.

If the intent is unclear, ask: "Are you looking to build/create evals, run them, or debug failures?"

Before Starting

Check memory for project-specific context (app ID, variable handling rules, audio scoring workarounds). If not available, ask the user.

© GoogleCloudPlatform, Apache-2.0. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 111 other files (scripts, references, assets) in .agents/skills/cxas-agent-foundry of GoogleCloudPlatform/cxas-scrapi.

  • SKILL.md
  • agents/coverage-analyst.md
  • agents/eval-writer.md
  • agents/lint-fixer.md
  • agents/scaffolder.md
  • agents/tdd-writer.md
  • agents/triage-failure.md
  • assets/project-template/README.md
  • assets/project-template/cxas_app/Sample_Support_Agent/agents/root_agent/after_model_callbacks/after_model_callbacks_01/python_code.py
  • assets/project-template/cxas_app/Sample_Support_Agent/agents/root_agent/before_agent_callbacks/before_agent_callbacks_01/python_code.py
  • … and 102 more

Open the folder on GitHubat commit ffba639

Compare with similar skills

Cxas Agent Foundry next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Cxas Agent Foundry compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Cxas Agent Foundry this skillGoogleCloudPlatform/cxas-scrapi106—~2.4kAutomated safety check: PassApache-2.0
Ad ReviewCorridorTech/PoseCap220—~2.4kAutomated safety check: NotesApache-2.0
Auto HarnessPacificStudio/openase268—~1.2kAutomated safety check: PassApache-2.0
Harness LintCorrectRoadH/OpenTickly306—~396Automated safety check: PassAGPL-3.0
CCPM Project Managementautomazeio/ccpm8.4k—~1.1kAutomated safety check: PassMIT
Qwen Agentthananon/9arm-skills3.2k—~1.5kAutomated safety check: PassNone

Similar skills

  • Ad Review

    CorridorTech/PoseCap

    Two-axis fresh-context code review per WORKFLOW §10. An agent skill from CorridorTech/PoseCap.

    220 GitHub stars~2.4k tokensUpdated today
    DevelopmentAuto-check: notes

  • Auto Harness

    PacificStudio/openase

    Diagnose and strengthen a repository's harness layer: AGENTS.md rules, knowledge layout, architecture boundaries, lint and type gates, API and generated-client contracts, test scaffolding…

    268 GitHub stars~1.2k tokensUpdated 1 mo ago
    DevelopmentAuto-check passed

  • Harness Lint

    CorrectRoadH/OpenTickly

    A skill your agent uses when working in this repository after receiving recurring coding feedback, adding or reviewing durable lint guardrails from AGENTS.md, updating rules/.md, or finishing…

    306 GitHub stars~396 tokensUpdated 5 days ago
    DevelopmentAuto-check passed

  • CCPM Project Management

    automazeio/ccpm

    Runs a spec-driven workflow from PRD to epic to GitHub issues to parallel agents, with status, standup and blocked-work reports from bundled scripts.

    8.4k GitHub stars~1.1k tokensUpdated 6 mo ago
    Product & Project ManagementAuto-check passed

  • Qwen Agent

    thananon/9arm-skills

    Delegate menial, well-scoped coding tasks to a cheap Qwen-backed subagent via the claude-9arm command instead of burning Claude tokens/quota.

    3.2k GitHub stars~1.5k tokensUpdated 3 mo ago
    DevelopmentAuto-check passed

  • Bench Baton

    nooga/let-go

    Coordinate heavy local workloads across worktrees, processes, and subagents — benchmarks and timing-sensitive gates run exclusively on a quiesced machine, while builds, test suites, regeneration…

    568 GitHub stars~821 tokensUpdated yesterday
    DevelopmentAuto-check passed

More from GoogleCloudPlatform/cxas-scrapi

All 14 skills in this repo

  • Cxas Autolabel Rules

    GoogleCloudPlatform/cxas-scrapi

    Author, validate, and manage Contact Center AI (CCAI) Insights Autolabeling Rules.

    106 GitHub stars~1.1k tokensUpdated yesterday
    Auto-check passed

  • Cxas Configurable Dashboards

    GoogleCloudPlatform/cxas-scrapi

    Author, validate, and manage Contact Center AI (CCAI) Insights Configurable Dashboards.

    106 GitHub stars~1.8k tokensUpdated yesterday
    Auto-check passed

  • Cxas Dfcx Migration

    GoogleCloudPlatform/cxas-scrapi

    Migrate Dialogflow CX (DFCX) agents to CXAS (Customer Experience Agent Studio) agents.

    106 GitHub stars~3.2k tokensUpdated yesterday
    Auto-check passed

  • Cxas Loss Analysis

    GoogleCloudPlatform/cxas-scrapi

    Retrieves non-contained CCAI Insights conversations (losses), uses agent intelligence to cluster them into common failure patterns, and generates a markdown report.

    106 GitHub stars~1.5k tokensUpdated yesterday
    Auto-check passed

  • Cxas Sim Eval

    GoogleCloudPlatform/cxas-scrapi

    Converts CXAS golden evaluations to SCRAPI SimulationEvals test cases.

    106 GitHub stars~1.2k tokensUpdated yesterday
    Auto-check passed

  • Cxas Composite Voice Agent Optimizer

    GoogleCloudPlatform/cxas-scrapi

    Audits, optimizes, and remediates CXAS agent configurations for Gemini Composite V1 voice naturalness, persona styling, and multi-language coverage directly in local workspaces with cxas-scrapi.

    106 GitHub stars~10k tokensUpdated yesterday
    Auto-check passed

Categories

Questions about Cxas Agent Foundry

What does Cxas Agent Foundry do?

End-to-end GECX/CXAS/CES conversational agent lifecycle -- build agents from requirements (PRD-to-agent), create and run evals (goldens, simulations, tool tests, callback tests), debug failures, and…. Cxas Agent Foundry is an agent skill from GoogleCloudPlatform/cxas-scrapi. End-to-end GECX/CXAS/CES conversational agent lifecycle -- build agents from requirements (PRD-to-agent), create and run evals (goldens, simulations, tool tests, callback tests), debug failures, and iterate to production quality.

When should I use Cxas Agent Foundry?

Cxas Agent Foundry fits situations like: the user mentions GECX; conversational agents; pushing/pulling/linting agents; agent instructions/callbacks/tools on the Google Customer Engagement Suite platform.

How do I install Cxas Agent Foundry in Claude Code?

Run `npx skills add GoogleCloudPlatform/cxas-scrapi --skill cxas-agent-foundry -a claude-code`. Or copy the skill folder (.agents/skills/cxas-agent-foundry in GoogleCloudPlatform/cxas-scrapi) into .claude/skills/cxas-agent-foundry in your project. Claude Code loads it when a task matches its description.

How do I install Cxas Agent Foundry in Codex?

Run `npx skills add GoogleCloudPlatform/cxas-scrapi --skill cxas-agent-foundry -a codex`. Or copy the skill folder (.agents/skills/cxas-agent-foundry in GoogleCloudPlatform/cxas-scrapi) into .agents/skills/cxas-agent-foundry in your project. Codex loads it when a task matches its description.

Can I use Cxas Agent Foundry in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add GoogleCloudPlatform/cxas-scrapi --skill cxas-agent-foundry -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/cxas-agent-foundry, .gemini/skills/cxas-agent-foundry, .github/skills/cxas-agent-foundry and .opencode/skills/cxas-agent-foundry in your project.

What does Cxas Agent Foundry need to run?

Going by SKILL.md and its folder, Cxas Agent Foundry needs Python for the scripts in its folder and the command-line tools its instructions call (python). Our summary lists: Python 3.

Does Cxas Agent Foundry access the network?

SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.

Is Cxas Agent Foundry safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. The check reads SKILL.md only: the scripts in the folder are not scanned, so read them before running anything.

What licence does Cxas Agent Foundry use?

Cxas Agent Foundry is published under the Apache-2.0 licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Cxas Agent Foundry use?

About 2.4k tokens (SKILL.md is roughly 9.5k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full. Its references folder adds about 50k tokens, read only when the agent opens those files.

What are the alternatives to Cxas Agent Foundry?

Skills that share tags, products or a category with Cxas Agent Foundry: Ad Review (CorridorTech/PoseCap, 220 stars), Auto Harness (PacificStudio/openase, 268 stars), Harness Lint (CorrectRoadH/OpenTickly, 306 stars) and CCPM Project Management (automazeio/ccpm, 8.4k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Cxas Agent Foundry?

GoogleCloudPlatform (a GitHub organization) maintains it in GoogleCloudPlatform/cxas-scrapi, which has 106 GitHub stars. The repository holds 14 skills in this directory. The repository was last updated on October 6, 2026.

Source: GoogleCloudPlatform/cxas-scrapi on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.